Contendas Mirante Belts (Bahia, Brazil)" Geologic and Isotopic Constraints on the Sources

Chemical Geology, 83 (1990) 325-338 325 Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands

The 2-Ga peraluminous magmatism of the Jacobina- Contendas Mirante belts (Bahia, Brazil)" Geologic and isotopic constraints on the sources

Pierre Sabatd 1, Moacyr M. Marinho2, Philippe Vidal 3 and Michelle Caen-Vachette 3 LORSTOM, Paris (France) and Instituto de Geoci~ncias, UFBA, 40161 Salvador (BA) (Brazil) 2Companhia Baiana de Pesquisa Mineral (CBPM), Centro Administrativo da Bahia, 41500 Salvador (BA) (Brazil) aCNRS and Blaise Pascal University, F-63038 Clermont-Ferrand (France) (Accepted for publication February 2, 1990)

ABSTRACT

Sabat~, P., Marinho, M.M., Vidal, Ph. and Caen-Vachette, M., 1990. The 2-Ga peraluminous magmatism of the Jacobina- Contendas Mirante belts (Bahia, Brazil): Geologic and isotopic constraints on the sources. In: B.K. Nelson and Ph. Vidal (Guest-Editors), Development of Continental Crust through Geological Time, Chem. Geol., 83: 325- 338.

A 500-km north-south alignment of granitic intrusions cross-cuts the central part of the Silo Francisco craton. These late- to post-tectonic granitic bodies are emplaced between two Archean blocks (Jequi~ and Gaviio blocks) and cross-cut the Contendas Mirante volcano-sedimentary sequence and the metasedimentary rocks of the Serra de Jacobina. They are two-mica or muscovite-garnet-bearing peraluminous granites. Rb-Sr systematics show that these granites were emplaced during the Transamazonian orogeny ( ~ 1.9 Ga). The high initial STSr/86Sr ratios (0.706-0.748) and the very low £'Nd(t) ( -- 13 to -5) indicate a crustal origin. The possible sources are: ( 1 ) the Jequi6 block; (2) the Contendas Mirante sequence; and (3) the Gaviio block, including some early Archean domes [trondhjemitie-tonalitic-granodioritic (TTG), Boa Vista type ] which were tectonically emplaced within the volcano-sedimentary sequence. The currently available Rb--Sr and Sm-Nd data are not consistent with reworking of the 3.5-Ga TTG. The combination of field and isotopic constraints preclude the Jequi~ block and the Contendas Mirante sequence and favour instead the Gaviio block medium-grade terrains as the source for the peraluminous granitic line. The available data suggest that a continent-continent collision occurred during the Transamazonian orogeny, which followed subduetion/obduction marked by volcanism (are-tholeiitic, calc-alkalineand shoshonitic) and related plutonism.

1. Introduction et al., 1981 ) which consolidated at 1.7 Ga is mainly exposed in Bahia state. The Sao Francisco craton (Almeida, 1967; Available geochronologic data (Marinho et Almeida et al., 1977) provides an excellent area al., 1979, 1980; Brito Neves et al., 1980; Cor- to study crust formation events that occurred dani et al., 1985 ) recently synthesized by Mas- during the transition between the Archean and carenhas and Garcia (1987), and supple- Proterozoic ( ~ 2.5 Ga) and the orogenic pro- mented (Wilson, 1987; Wilson et al., 1988) cesses that operated during the lower-middle indicate a remarkable succession of igneous Proterozoic ( ~ 2 Ga). The Sao Francisco era- events from 3.5 to 1.9 Ga. In addition, a long ton is one of the main remnants of the Archean alignment of 2-Ga-old peraluminous leuco- and early Proterozoic crust of South America. granites (Himalayan-type) marks the limit be- The part of this structural province (Almeida tween the Jequi6 block and the Gavi~o block.

This boundary is also marked by low- to me- sedimentary sequence. In the Atlantic Coast dium-grade supracrustal fold belts. The exis- belt (Fig. 2), Barbosa (1986) presented evi- tence of good indicators of a continent-conti- dence for volcanic series similar to modern is- nent collision, such as high-grade land arc associations. The distribution of the metamorphism (Newton, 1987) and leuco- arc tholeiites, and the calc-alkaline and the granites (Le Fort, 1981 ), in addition to the ex- shoshonite metavolcanics of these series sug- istence in the Jequi6 block of a lower Protero- gests an early subduction system preceding the zoic arc series metamorphosed in the granulite strong deformation and granulitic metamor- facies (Barbosa, 1986), indicates that a conti- phism. This subduction system is also recog- nent-continent collision occurred around 2 Ga nized by Figueiredo (1989) on the basis of following subduction under the Jequi6 block. geochemical data. The goals of the present work were: ( 1 ) to determine the absolute chronology of the mag- 2.2. Medium-gradeterrains matic episodes; (2) to identify the sources of The medium-grade terrains occupy roughly the different magmas; and (3) to constrain the the western part of the S~o Francisco craton geodynamic processes that occurred during the and correspond to the Gaviao block (Fig. 1 ). Transamazonian orogeny. The separated nucleus of the middle Itapicuru 2. Geologic setting river in the northeastern part of the craton, now juxtaposed with the Salvador Cura~a granulit- The main Archean to Proterozoic features of ic belt, has similar lithologies and metamor- the S~o Francisco craton in Bahia state, Brazil, phic history, and is related to the Gavi~o block. may be summarized by three commonly ac- The terrains are composed of an assemblage cepted units (Fig. 1 ). of various gneiss-amphibolite associations, (1) High-grade terrains form the Jequi6 which include migmatites and plutonic rocks. block and its mobile belts (Mascarenhas, 1973, Some of the gneiss formations are paragneiss, 1979) extend for >700 km from north to but most of the gneiss-leptite-amphibolite south, and correspond to the Salvador Curaqa piles are considered as volcanic and volcano- and the Atlantic Coast or Itabuna granulitic clastic or volcano-sedimentarysequences. The belts, respectively (Fig. 2). plutonic rocks are widely distributed in the (2) A medium-grade gneiss-migmatitic gneiss migmatite complex. Among them a 3.1- complex forms the large western band of these 3.5-Ga trondhj emitic-tonalitic-granodioritic older terrains (Gavi~o block) and a northeast- (TTG) association forms huge massifs such as ern nucleus in the middle Itapicuru river Lagoa do Morro, Sere Voltas, Boa Vista, etc. region. (Fig. 2 ), which probably represent the oldest (3) Supracrustal sequences, metamor- lithologies of the craton (Cordani et al., 1985 ). phosed in the greenschist to amphibolite fa- The rocks were deformed by tectonic events cies, are associated with the medium-grade during the Archean and Proterozoic. The lat- gneiss migmatitic complex (Fig. 2). They cor- ter is mainly marked by E-W shortening, as- respond to the volcano-sedimentary belts of sociated with westward thrusting and crustal Jacobina and Contendas Mirante as well as less thickening related to the last migmatitic pro- extensive occurrences. cesses (Sabat6 et al., 1988 ).

2.2. High-grade terrains 2.3. Supracrustal sequences

The high-grade terrains consist of granulite- The most representative supracrustal se- facies charnockites, enderbites, and a volcano- quences are located along the junction be- THE 2-Ga PERALUMINOUS MAGMATISM OF THE JACOBINA-CONTENDAS MIRANTE BELTS 327

-- BA HIA t~, t~ AMBD c 9 40 ° 42 ° I

46 ° ~10 o

I0°--~ • ~%*" = /jl %

sb / / a

)" I ~ 12 ° "[2o .,-" / I \

/ \ ALVADOR /,,, "5( J ( 140 ",4 14°-.-..-.- \ \. ~x I I // / L // I I 46 o 44 ° e~",94 ,,# C5 __16 ° ~-~ V 42 °

~3 0 100 200Kin t I I 3 1 18°---- 1"~.,/ - __18o I 40 o

Fig. 1. Structural outline of Bahia state. [ 1 = Phanerozoic cover; 2 = Brazilian and Pre-Brazilian covers (SAo Francisco and Espinhago super groups); 3=Archean to early Proterozoic supracrustal complexes (volcano-sedimentary and greenstone belts); 4=Archean to early Proterozoic mobile belt (granulites, charnockites, migmatites and gneiss); 5=Jequi6 granulitic complex (granulite facies volcano-sedimentary cover, charnockites and enderbites); 6= Archean gneiss-migmatitic and granitic complex]. Modified from Mascarenhas (1976). 328 P. SABATI~ET AL

-~-9°38 °

Campo Forrnoso

Carnaiba

-- 11 o

......

:.-::, . .::..:: 0 501Kin / / /

SALVADOR

% Riacho das ,:.::

,. ...,,::..::

"**4 ".: *.'. .'.'.:' :--:::.: : :::.,~

~..~• °,~. :-~.x::.x.:.-::~!:!-.:..-::::: Lagoa do?i::.-::::-i~:~.....: " " :""" ..: ....t:... '"" :. ~~ ..-..:-.'"". -f---~l= " ...... • t.:....:.',.....;.'.:;....'...... :" ...... "~ ii:-:":".'-.~-:"::...;.,..., ...... '...... " ..... /...... '..'..." ":~ ::~-'.%::~.:i;:i::: :?:.!.:%-;. ~-:...--:.:;!.!-iii~~:ii: ii ii ii iii: :iii i: :~. :i i: !!!iii • ...... :.:::::::-!:,...... :::.::... ,,':..-:,...::::::..:.:::::-..:-,..::~=-

Fig. 2. Tectonic sketch map of the junction zone of the Oavi~o block and the Jequi~ block. Location of the studied intrusions relative to the main structures. 1= Phanerozoic and middle and upper Proterozoic covers and belt; 2 = volcano- sedimentary belts; 3 = high-grade terrains; Jequi6 block granulitic formations and Archean to early Proterozoic mobile belts; 4 = Gaviao block medium-grade formations; the late- to post-tectonic Transamazoniangranitoids are dark', the older (Archean) and syn-tectonic (early Transamazonian event) plutons are shown by a cross. THE 2-Ga PERALUMINOUS MAGMATISM OF THE JACOBINA-CONTENDAS MIRANTE BELTS 329 tween the two Archean blocks and form the Jacobina belt from the Salvador Curaga mo- main meridian alignment of Contendas Mir- bile belt (Figs. 1 and 2 ). ante and Jacobina belts (Figs. 1 and 2 ). Some remnants of sequences somewhat analogous to 3. The plutonic rocks the former are encountered into the Gavi~o block (Fig. 1 ). Each of these belts is associated All these belts are intruded by several plu- with plutonic rocks. tonic associations which are described as be- longing to two distinct groups. 2.3.1. Contendas Mirante belt. The volcano- The Jacobina and Contendas Mirante belts are the locus of a remarkable up to 500-km sedimentary sequence of Contendas Mirante is composed of two stratigraphic units (Marinho alignment of syn- to late-tectonic granites that et al., 1979, 1980). The lower unit contains al- emphasize the axial shear zone and the bor- ders of the belts. More than fifteen bodies are most all the volcanogenic components. These are basalts and intermediate metavolcanics in- recognized as related to the Contendas Mir- ante belt (Marinho et al., 1979, 1980; Mar- tercalated with detrital and chemical sedi- inho and Sabatr, 1982; Concei~o, 1986) and ments. The upper unit is essentially detrital. up to seven related to the Jacobina belt (Couto, The rocks are metamorphosed in the greenschist facies grading to the amphibolite facies 1978; Rudowski et al., 1987; Celino and Sa- at the proximity of its eastern tectonic contact batr, 1988). Some bodies can also be seen in with the Jequi6 block. The two units are de- the same tectonic alignment between the two formed by two principal phases (Sabat6 et al., belts (Fig. 2) in the Itaberaba region (Fer- 1980). The first is isoclinal recumbent, with a nandes and Sabatr, 1987; Giuliani et al., west vergence and is accompanied by a west- 1988 ). All of the plutons are peraluminous and ward overthrusting. The second is marked by show various degrees of differentiation (Con- E-W shortening which upturns the previous ceig~o, 1986; Rudowski, 1989 ). structures and progressively transforms the Seven areas were chosen for the present work: westward tangential movement into sub-horizontal N-S transcurrent shearing. With these -Five are located in the axial zone of the phases, segments of the Gaviao block are Contendas Mirante belt. From north to south uplifted in the Contendas Mirante belt and its they are the Gameleira, Riacho das Pedras, imbricated Gavi~o block segments. Caetano, Lagoa Grande and Lagoinha bodies. -The composite Campo Formoso massif and the Carnaiba body are situated in the Ja- 2.3.2. The Jacobina belt. the metasedimentary cobina belt. Jacobina sequence is mainly made up of detrital epicontinental sediments. Some ultramaf- 3.1. Gameleira massif ics are intercalated in the tectonic edifice. The rocks are metamorphosed in the amphibolite The Gameleira is the largest peraluminous facies. The structure of the belt corresponds to body of the axial line intrusions of the Conten- imbricated slices which also show westward das Mirante belt (Fig. 2 ). In the vicinity of the vergence. As in the Contendas Mirante belt, the Jequi6 block, it intruded the volcano-sedimen- subsequent shortening and upturning led the tary sequence where it induced a thermal tangential movement to be transformed to sub- metamorphism that developed cordierite and horizontal N-S shearing. In the resulting andalusite in the country rock. Its emplace- framework, a large overthrusted imbricated ment is synchronous with the late and impor- slice related to the Gaviao block separates the tant overprinting folding phase (Sabat6 et al., 330 P. SABATI~ET AL.

1980) which represents shortening related to with biotite (Caetano), biotite + garnet (La- the crustal thrusting. goa Grande), and biotite+muscovite (La- The mean composition is granodioritic with goinha). Aplopegmatitic dikes or pockets with a quartz-oligoclase-andesine-microcline as- biotite, muscovite and tourmaline are semblage showing wide variations in relative associated. proportions of the phases. The index mineral is a red-brown, green or sometimes bleached 3.4. Campo Formoso massif and Carnaiba biotite. Muscovite may be abundant. It is gen- body erally deuteric, formed at expense of the K- feldspar and plagioclase. Microcline is also present. The accessory minerals are zircon, The Campo Formoso massif is the most im- apatite, titanite, magnetite-ilmenite and sec- portant intrusion in the Jacobina belt. It rep- ondary epidote. resents a multi-stage intrusion with several successive, roughly concentric, peraluminous 3.2. Riacho das Pedras intrusion granitic facies, and a suite of aplite and pegmatite dikes (Rudowski et al., 1987). Three The Riacho das Pedras granite (Fig. 2 ) is late successive main types are recognized (Ru- tectonic relative to the folding phases and the dowski, 1988): (1)a coarse-grained to por- accompanying overthrusting. It consists of a phyritic muscovite-rich granite; (2) a me- hololeucocratic granite associated with a net- dium- to fine-grained two-mica granite and (3) work of aplite and pegmatite dikes that cross- a muscovite-garnet-albite granite with only cut the sequence. The fine- to medium-grained relicts of biotite. In these three types, zircon granite is made up of quartz, albite and mi- and apatite are the common accessories. Gar- crocline. Muscovite may be abundant. The net, tourmaline and beryl characterize the as- biotite is subordinate and sometimes absent. sociated quartz-albite-microcline aplopeg- Apatite, magnetite, hematite, zircon and rare matite suite. titanite are the typical accessory minerals, with Muscovite crystallized during several stages less commonly tourmaline and garnet. The two of the magmatic evolution. The early crystalli- latter minerals, especially tourmaline, are zation developed during the stage correspond- common in the related pegmatites and aplites. ing to the acquisition of the preferred orienta- tion by the plutonic bodies. The muscovite also 3.3. Caetano, Lagoa Grande and Lagoinha formed during the deuteric stage, and accom- bodies panies late shear zones or fills microfissures. The Carnaiba body is a small circular intru- The Caetano, Lagoa Grande and Lagoinha sion composed of two types of granites corre- intrusions form elongate bodies (20, 10 and 4 sponding to types (2) and (3) of Campo km long, respectively) aligned along the axial Formoso. shear zone of the Contendas Mirante belt. A Chemically, all the granitic facies are peral- narrow band of supracrustals separate them uminous. Based on trace-element and rare- form the large N-S Archean dome of Sete Vol- earth element (REE) behaviour, Rudowski tas (Fig. 2). Their emplacement is synchron- ( 1988 ) distinguishes two "series": one corre- ous (Lagoinha, Lagoa Grande) to late (Cae- sponding to the two-mica granite (type 2) tano) with respect to the shearing. which leads by differentiation to the musco- The plutons consist of leucogranite with a vite-garnet granite, and the other to a suite of two-mica +_ garnet association. They have muscovite granites (type I ) marked by subsol- muscovite layering (Caetano), and schlieren idus transformations. The chemical character- THE 2-Ga PERALUMINOUS MAGMATISM OF THE JACOBINA-CONTENDAS MIRANTE BELTS 3 31 teristics suggest a heterogeneous magmatic mations than in either the Gavi~o block or the source (Rudowski, 1989). Jequi6 high-grade lithotypes. The structural framework clearly shows the 4. Geologic constraints on the sources presence of Gavi~o block terrains as infras- tructure to the Contendas Mirante edifice. The The structural evolution indicates that per- geologic configuration of the leucogranitic in- aluminous plutonism accompanied conti- trusion line located on the western border of nent-continent collision marked by over- the Gavi~o block segment uplifted into the thrusting and subsequent shortening that Contendas Mirante belt, and their chronologic terminated with transcurrent shearing. relation as syn- to late-tectonic relative to the The presence in the Gameleira massif of transcurrent shearing following overthrusting, mica-rich garnet-bearing xenoliths, which may rule out the possibility that the Jequi6 block is represent the residual phase of partial melting, a good candidate for the source of the granites. led Petta (1979) to attribute the source rocks In addition, the granulitic metamorphism, to the neighbouring blocks. The overthrusting which is related to continental collision (New- of the Jequi6 block previous to or synchronous ton, 1987), and, in the case of the Jequi6 block, with the intrusion rules out the Jequi6 block is approximately synchronous with leucogran- alternative for the source, and restricts the ire intrusion (Wilson, 1987; Wilson et al., source to within the Gavi~o block formations 1988), indicates that the Jequi6 block should and/or the volcano-sedimentary sequence. separated in space from the peraluminous For the Riacho das pedras granite, as for the granite production zone. Thus the potential similar bodies of the axial Contendas Mirante sources probably are in the non-granulitic for- line (Caetano, Lagoa Grande, Lagoinha), a mations of the Gaviao block. Similarly, the crustal source is assumed by Marinho and Sa- chronologic and tectonic setting of the peralu- bat6 ( 1982 ) on the basis of the peraluminous minous diapirs of Campo Formoso and Car- and sodi-potassic paragenesis, and confirmed naiba, and their field relations with the me- by the very high 87Sr/86Sr ratio ( ~ 0.768) ob- dium-grade gneiss-migmatites related to the tained during a previous dating program (Sato, Gavi~o block, support the Gaviao terrains as a 1982; Cordani et al., 1985 ). Compared with the possible source and rule out involvement of the Gameleira intrusion, they probably had the granulitic Jequi6 block. The deep metasedi- same petrogenesis. Crystallization occurred mentary Jacobina sequence, intruded by the under conditions corresponding to the meso- granites may also have contributed to their zone and epizone where mica and cordierite are genesis. in equilibrium, implying a high water pres- Are the supracrustals a possible source for sure. Thus, the rocks must have crystallized the granites? The lower unit of the Contendas relatively close to their source regions. How- Mirante sequence is composed of ultramafics, ever, the migration of leucogranitic magmas mafics, carbonate rocks and banded iron for- could have been more extensive because of mations, and consequently is not a good can- higher fluid pressure linked to the presence of didate. The clastic sediments, tufts and gray- boron, which reduces both the solidus temper- wacke of the upper unit are a possible source; ature and the magma viscosity (Pichavant, however, they are cross-cut by the granites and 1984, 1987). This is suggested as much by the are located in the emplacement zone of the petrographic associations, with its abundant massifs, not in the source zone of the magma. pegmatite suites, as by the presence of tour- The total absence of mafic enclaves and the maline. Consequently, the source is less likely geochemical patterns (Rudowski, 1989; Cu- to have been in the Contendas Mirante for- ney et al., 1990) tend to exclude a mantle 332 P. SABATI~.ET AL. source. The composition of all the granites, TABLE I which is very close to the ternary minimum of Rb-Sr isotopic data for Bahia Transamazonian peralu- the granitic system, and the absence of frac- minous granites tionated crystallization suites (Cuney et al., Sample Rb Sr 87Rb/86Sr 878r/86Sr 1990) support the hypothesis of magma pro- (ppm) (ppm) duction exclusively by crustal melting. Alter- natively, Cuney et al. (1990) show that the Campo Formoso granite: mineralogical and chemical features of these 6 230 167 4.040 0.82143 granites reflect the dominantly meta-igneous 13 265 153 5.067 0.85094 22c 265 129 6.058 0.87784 character of the source and propose the Ar- 180 320 121 7.808 0.92704 chean TTG as the possible source material. 88 289 105 8.148 0.93140 54A 274 93 8.697 0.94536 5. Isotopic study 29 285 89 9.542 0.98643 62A 371 71 15.90 1.16032 48C 282 30.3 29.050 1.52168 5.1. Analytical techniques 48B 342 37.0 28.847 1.50898

Carnaibagranite: Sr and Nd isotopic compositions have been obtained in Clermont-Ferrand, on a VG54E ® CA5 350 90.7 11.53 1.04894 CA9 404 92.0 13.16 1.08494 mass spectrometer using the procedures de- CA16 460 64.0 22.08 1.33340 scribed by Pin and Carme ( 1987 ). During the course of the experiments, NBS 987 and La Riacho das Pedras granite: Jolla standards gave 875r/86Sr and 143Nd/144Nd 99D 215 13.5 53.03 2.25953 of 0.71025_+0.00004 and 0.51186+_0.00003, 99H 281 6.0 211.9 6.53939 991 134 14.5 28.89 1.54825 respectively (the quoted errors represent the 99J 334 5.5 349.6 10.48640 external reproducibility at 2aon 10 Sr runs and 99K 311 5.6 291.2 8.96639 8 Nd runs). Sm and Nd were obtained by iso- 99B 200 14.5 44.68 1.96824 99(2 215 14.4 49.07 2.08941 tope dilution with a precision of -+ 1% on their 99F 214 15.5 44.89 2.01100 ratio. Rb and Sr were obtained by X-ray fluo- 99M 327 4.8 417.8 12.22670 rescence spectrometry (XRF) with a precision of -+ 2% on their ratio. Game&iragranite: 91A 271 146 5.429 0.86052 92 225 172 3.828 0.81478 5.2. Results 95 282 161 5.129 0.84709 96 259 94 8.124 0.93353 Five granites were dated using the Rb/Sr 97B 276 146 5.582 0.86737 101 275 114 7.114 0.90776 whole-rock isochron technique. The data are 11218 342 148 6.790 0.89570 reported in Table I and in Fig. 3. Three of the granites (Campo Formoso, Gameleira and Lagoa Grande-Lagoinha-Caetano granite: Riacho das Pedras) yield well-defined iso- GO9HC-BC 243 76 9.430 0.97237 chrons and cluster in the age range of 1.93-1.97 G54HC 205 182 3.300 0.80295 Ga. The isochrons for the two other granites G58HC 173 248 2.028 0.76428 CAE2 66.0 31.0 6.241 0.84683 are based only on three samples and need con- firmation, although it is probably not fortui- tous that the age of the Lagoinha-Lagoa reflect disturbance during the Brazilian event. Grande granite lies within the same range as In addition, the Carnaiba granite is very prob- the others. The Caetano data plot below the ably the source of fluids that developed the Lagoinha-Lagoa Grande isochron, and may emerald-rich metasomatic zones in the enclos- THE 2-Ga PERALUMINOUS MAGMATISM OF THE JACOB1NA-CONTENDAS MIRANTE BELTS 33 3

1.6 - 87Sr/ 86Sr 87Sr 86Sr .~/ 1.5 o.9ooh- ./J : CAMPO FORMOS~ 1.4 1.3

1.2 1F (y/ t = 1947 +- 54 Mo | ./w ISr =0.7072 _+ 0.0038 I.I / t = 1969+29 Mo (.y,,,,,/ "rSr = 0.7062 + 0.0028 1.0 MSWO = 3.2 0.90C 0.750 hF I" 0,80C S 87Rb/ 86Sr I J 87Rb/86Sr 0.70C I I I i I 0.700 r i i i i t i i i i 5 I0 i5 20 25 30 0 I 2 3 4 5 6 7 8 9

87Sr/ 86Sr 1.0 87Sr/ 86Sr LAGOINHA RIACHO DAS P~~ 0.90C LAGOA GRAND~.

I0 /~" t = 1929 + 16 Mo Isr=0748_+ 0013 080(: MSWD = 13 ~- MSWD = 2,9 87Rb/86Sr J 87Rb/ 86Sr I 0.700 I I I I IO0 2C)0 3~)0 400 ' 500I 2 4 6 8

1,4 aTSr/ 86Sr 1.5 CARNAIBA ~ , 1.2

I.I

1.0 MSWD = 08

0.900 0.800 / a'Rb/ e6Sr 0.700 Fig. 3. Rb-Srisochron diagrams for the BahiaTransamazonian peraluminous granites. Is, = initialSr ratio= (87Sr/86Sr)~. ing serpentinites (Rudowski, 1989). Rb-Sr subsequent to the main phase of the Transa- dating of these zones provided an age of mazonian emplacement between 1.97 and 1.93 1.869 _ 0.028 Ga (L. Rudowski, Ph. Vidal and Ga. M. Caen-Vachette, in prep.) which is within The 87Sr/86Srinitial ratios of the granites are error of the 1.88-Ga age obtained on the Car- much higher than the mantle ratio at the time naiba granite. Therefore, the Camaiba granite of their production (Fig. 4). This indicates that probably represents a late-magmatic event they were formed by melting of materials hav- 334 P. SABATI~ ET AL.

MANTLE RIACHO DAS PEDRA~ .~ GAMELEIRA -~ CAMPO FORMOSO o CARNAIBA --0-- LAGOA GRANDE- LAGOINHA

GAVI~O BOA VISTA-MATA VERDE__ LAGOA DO MORRO JEQUI~ --CONTENDAS MIRANTE-- __CONTENDAS MIRANTE (LOWER UNIT ) 0.710 (UPPER UNIT) 0.730 0.750 87^r-86.-,;~/~ I ] I r

Fig. 4. 87Sr/86Sr vs. time for Bahia Transamazonian granitoids and their potential source rocks. The field of the source rocks have been drawn using the data of Cordani et al. ( 1985 ) and M.M. Marinho (unpublished data, 1988 ). ing had a significant crustal residence time. To proposed to result from exchange with country consider potential sources, the range of 87Sr/ rocks through convective cells (Bonin et al., 86Sr ratios (corrected for 1.95 Ga and using 1987 ). However, there is no indication of hy- their extreme Rb/Sr ratios and isotopic com- drothermal activity or low-temperature position) of the various formations exposed in changes in granite mineralogy. this part of the Sao Francisco craton have been In conclusion, Sr isotopes alone cannot plotted in Fig. 4. The field for Boa Vista-Sete identify with precision the source rocks. In ad- Voltas and Lagoa de Morro TTG, the Gaviao dition, it must be recalled that despite the fact block migmatites and granitoids, and the Je- that Sr isotope geochemistry indicates that all qui6 high-grade rocks are rather well con- granitoids possess a crustal signature, the exact strained (Cordani and Iyer, 1978; Cordani et characteristics of their sources are impossible al., 1985; Wilson, 1987). In contrast, the Rb- to deduce by Rb-Sr systematics since variable Sr characteristics of the Contendas Mirante Rb/Sr ratios can be produced by partial melt- volcano-sedimentarysequence are poorly con- ing of a given source without changing 87Sr/ strained. We have reported in Fig. 4 the few 86Sr" available data for the formations which could, In contrast to Sr isotopes, Nd isotopes are from petrologic arguments, possibly be good more straightforward to interpret because Sm/ candidates for the granitic source, i.e. the phyl- Nd ratio fractionation during partial melting lites of the upper unit and the felsic metavol- is much smaller and more confidently mo- canics of the lower unit (Cordani et al., 1985; delled. The Sm-Nd results are given in Table Wilson, 1987). The Gameleira, Campo For- II. The calculated ENd(t)-values are negative moso and Lagoa Grande-Lagoinha granites ( - 13 to - 5) and confirm the crustal origin of match only the Jequi6 and Gavi~o fields, the granitoids as inferred from Sr isotopes. The whereas the origin is less constrained for the time of first extraction of the material out of Riacho das Pedras and Carnaiba granites. the mantle can be obtained by calculating However, for the Riacho das Pedras granite, its model ages relative to the depleted mantle. very high initial Sr isotopic ratio (0.747) might Ages range between 2421 and 3165 Ma. The not be that of the source because this highly case of Riacho das Pedras granite (tDM= 3165 evolved magma had exceedingly high Rb/Sr Ma is peculiar. It exhibits very high Rb/Sr ra- ratios. The 87Sr/S6Sr ratio can increase very tios, low REE contents and high Sm/Nd ra- rapidly during the magmatic stage itself (Vidal tios. These features are typical of highly et al., 1979). High initial ratios also have been evolved granites. In this case, the Sm/Nd ratio THE 2-Ga PERALUMINOUS MAGMATISM OF THE JACOBINA-CONTENDAS M1RANTE BELTS 335

TABLE II

Sm-Nd isotopic data for Bahia Transamazonian peraluminous granites tDM is calculated using (143Nd/~44Nd)~=0.513114 and :47Sm/144Nd=0.222 (Ben Othman et al., 1984)

Sample Granitoids Sm Nd 147Sm/144Nd t43Nd/144Nd (Nd(O IDM (ppm) (ppm) (Ma)

11217 Riacho das Pedras 0.55 2.02 0.1646 0.51191 -6.2 3,165 11218 Gameleira 4.1 27.2 0.09111 0.51087 -8.8 2,600 GO9HC-BC Lagoinha 4.1 17.5 0.1420 0.51169 -4.9 2,698 G58HC Lagoa Grande 2.5 15.6 0.0966 0.51097 - 7.6 2,592 CAE2 Caetano 1.47 10.03 0.0889 0.51099 -5.2 2,421 6 Campo Formoso 5.81 41.1 0.08536 0.51057 - 13.1 2.821 13 Campo Formoso 4.07 25.5 0.09646 0.51101 -7.4 2,541

~Nd CONTENDAS MIRANTE UPPER UNIT ~r,u \

-2

2'5 Fig. 5. Nd vs. time for Bahia Transamazonian granitoids and their potential source rocks. The field of the source rocks have been drawn using the data of Wilson (1987) and M.M. Marinho (unpublished data, 1988 ). of the magma can be increased with respect to ages (Vidal et al., 1984; Vidal, 1987). Instead, the source by retaining monazite in the partial using a typical crustal REE pattern (e.g., shale melting residue, or removing it during an early pattern) for the source provides a source age stage of the magmatic history. This situation in agreement with the other granitoids. can lead to a gross overestimation of the model The age of the sources, the close grouping of 336 P. SABAT~ETAL

IDM between 2.4 and 2.8 Ga, and comparison topes are far less sensitive to this effect. This is of the l~Nd(t) of the granites with their potential reflected in the Campo Formoso granite where sources (Fig. 5) (data from Wilson, 1987; the two analyzed samples have different M.M. Marinho, unpublished data, 1988), ENd(l). strongly support the hypothesis that the crust that has produced these granitoids by partial Acknowledgements melting is neither the 3.5-Ga-old Boa Vista- The results presented in this paper consti- Mata Verde orthogneisses nor the 3.1-Ga-old tute a contribution of the CNRS-INSU pro- Lagoa do Morro orthogneisses (which origi- gram "Dynamique et Bilan de la Terre" and of nally belonged to the Gavi~o block), Instead the project "Granit6ides da Bahia: geologia e the sources must be late Archean in age. metalog~nese" which receives financial support through agreements FINEP (PADCT) / 5.3. Discussion UFBA, CNPq/ORSTOM, SGM-SME (Ba)/ UFBA and CNEN/UFBA. Although Nd isotopes are consistent with the The authors thank J. de F. Mascarenhas felsic rocks from the Contendas Mirante se- (SGM-SME Bahia) for helpful discussions in quence (upper and/or lower unit) as the the field and L. Rudowski for donating the source of the Transamazonian peraluminous samples of the Campo Formoso and Carnaiba granites (Fig. 5 ), the hypothesis is ruled out by massifs. We also especially thank B. Auvray, structural constraints. The same conclusions H. Martin, B. Nelson and three anonymous re- are reached for the Jequi6 block formations. viewers for useful criticism and comments and In addition, isotopic data also show that the I. McReath for helpful discussions and revi- old TTG cannot be considered as a good can- sion of the English text. didate for the source. Therefore, the only possible source left is the late Archean migma- References tized gneiss-amphibolitesand granitoids of the Gavi~o block. These fit the structural, petro- Almeida, F.F.M. de, 1967. Origem e evolu~o da plata- forma brasileira. Dep. Nac. Prod. Min., Div. Geol. logic and Sr isotope constraints. Although there Min., Bol. 241, 36 pp. are no Nd isotopic data available yet, at the Almeida, F.F.M. de, Hasui, Y., Brito-Neves, B.B. de and present time this is the best working hypothesis. Fuck, R.A., 1977. Provincias estruturais brasileiras. 8 ° Finally, in the attempt to relate the produc- Symp. Geol. Nordeste, Campina Grande, Atas, Soc. Bras. Geol., pp. 363-391. tion of the peraluminous granites to a conti- Almeida, F.F.M. de, Hasui, Y., Brito-Neves, B.B. de and nent-continent collision, the similarity in the Fuck, R.A., 1981. Brazilian structural provinces: an isotopic characteristics of the Transamazon- introduction. In: J.M. Mabesoone, B.B. de Brito-Neves ian and Himalayan granites is striking. In both and A.N. Sial (Editors), The Geology of Brazil. Earth- Sci. Rev., 17: 1-130. cases, very low ENd(/)-values are observed Barbosa, J.S., 1986. Constitution lithologique et mrta- which implies essentially a crustal contribu- morphique de la rrgion granulitique du sud de Bahia, tion. The only difference is the apparent iso- Brrsil. Doct. Thesis, University of Paris VI, Mrm. Sci. topic homogeneity in the Transamazonian Terre, Paris, No. 86-34, 401 pp. granites suggested by the good straight line of Ben Othman, D., Fourcade, S. and All~gre, C.J., 1984. Recycling processes in granite-granodiorite complex the data points in the Rb-Sr isochron dia- genesis: the Querigut case studied by Nd-Sr isotope grams. This differs strongly from the Himala- systematics. Earth Planet. Sci. Lett., 69: 290-300. yan situation (Vidal et al., 1982; Deniel et al., Bernard-Griffiths, J., Peucat, J.J., Sheppard, S.M.F. and 1987 ). This could simply be due to an "aging Vidal, Ph., 1985. Petrogenesis of Hercynian leucogranites from the Southern Armorican Massifs: Con- effect" (Vidal et al., 1985 ) which tends to wipe tribution of REE and isotopic (Sr, Nd, Pb, O) geo- out the initial isotopic heterogeneities. Nd iso- chemical data to the study of source rocks THE 2-GaPERALUM1NOUS MAGMATISM OF THE JACOBINA-CONTENDASMIRANTE BELTS 337

characteristics and ages. Earth Planet. Sci. Lett., 74: cations. VII Congr. Lat. Am. Geol., Bel6m, Abstr. 235-250. Vol., p. 349. Bonin, B., Platevoet, B. and Vialette, Y., 1987. The geo- Le Fort, P., 1981. Manaslu leucogranite: a collision sig- dynamic significance of alkaline magrnatism in the nature of the Himalaya, a model for its genesis and Western Mediterranean compared with West Africa. emplacement. J. Geophys. Res., 86(Bll): 10545- Geol. J., 22: 361-387. 10568. Brito-Neves, B.B. de, Cordani, U.G. and Torquato, J.R.F., Marinho, M.M. and Sabat6, P., 1982. The Contendas- 1980. Evolu~o geol6gica do Precambriano no Estado Mirante volcano-sedimentary sequence and its grani- da Bahia. In: H.A.V. Inda (Editor), Geologia e recur- tic-migmatitic basement. In: International Sympo- sos minerais do Estado da Bahia, Vol. 3. CPM/SME sium on Archean and Early Proterozoic Geologic evo- (Coord. Prod. Miner./Secr. Minas Energ. ) Bahia, Sal- lution and Metallogenesis, Vol. 1, CPM/SME (Coord. vador, pp. 1-10 h Prod. Miner./Secr. Minas Energ.) Bahia, Salvador, Celino, J.J. and Sabat6, P., 1988. O Maciqo de Jaguarari: Abstr. Excurs., pp. 139-184. uma intrus~o granitica de vocaq~o crustal na margem Marinho, M.M., Silva, E.F.A. da, Soares, J.V. and Costa, do cintur~o transamaz6nico da Serra de Jacobina P.H., 1979. Projeto Contendas-Mirante. Comp. Baiana (Bahia, Brasil). VII Congr. Lat.-Am. Geol., Bel6m, Pesqui. Miner. Conv. CPBM (Comp. Baiana Pesqui. Abstr. Vol., p. 437. Miner. )/SME (Secr. Minas Energ. ) Bahia, Salvador, Concei~o, H., 1986. Os granitos do Rio Caveira: Petro- Relat6rio Fin., Vol. 1 (unpublished). logia de intrus6es Precambrianas no cisalhamento ax- Marinho, M.M., Silva, E.F.A. da, Lopes, G.A.D.C., Soares, ial do complexo Contendas Mirante (Bahia-Brasil). J.V. and Cruz, M.J.M., 1980. Projeto Anag6-Caldei- Dissertation, Mestrado, Univ. Feder. Bahia, Salvador rao. Comp. Baiana Pesqui. Miner. Conv. CBPM/SME (unpublished). (Comp. Baiana Pesqui. Miner./Secr. Minas Energ.) Cordani, U.G. and Iyer, S.S., 1978. Geochronological in- Bahia, Salvador, Relat6rio Fin., Vol. 1 (unpublished). vestigation on the Precambrian granulitic terrains of Mascarenhas, J. de F., 1973. A geologia do centro-leste da Bahia, Brazil. Precambrian Res., 9: 255-274. Bahia. An. 27 Congr. Bras. Geol., Aracaju, Soc. Bras. Cordani, U.G., Sato, K. and Marinho, M.M., 1985. The Geol., 2: 35-66. geologic evolution of the ancient granite-greenstone Mascarenhas, J. de F., 1976. Geologia da region centro terrane of central-southern Bahia, Brazil. Precambrian oriental da Bahia: Projeto Bahia, II. Sul da Bahia. Res., 27: 187-213. CPRM, Conv. DNPM/CPRM (Dep. Nac. Prod. Couto, P.A., 1978. Projeto Serra de Jacobina. Relat6rio Miner./Comp. Pesqui. Recurs. Miner. ), Salvador, Re- Fin., DNPM-CPRM (Dep. Nac. Prod. Miner.-Comp. lat6rio Fin., 1 vol., maps. Pesqui. Recurs. Miner.), Salvador, Vol. 1. Mascarenhas, J. de F., 1979. Evolug~o geotect6nica do Cuney, M., Sabat6, P., Vidal, Ph., Marinho, M.M. and Precambriano do Estado da Bahia. CPM/SME Concei~ao, H., 1990. The 2 Ga peraluminous mag- (Coord. Prod. Miner./Secr. Minas Energ.) Bahia, matism of the Jacobina-Contendas Mirante belts Geol. Rec. Min. Est. Bahia, Salvador, Textos B~isicos, (Bahia, Brazil): major and trace elements geochemis- 2: 57-165. try and metallogenic potential. In: P. Le Fort (Guest Mascarenhas, J. de F. and Garcia, T.W., 1987. Mapa geo- Editor), Collision Magmatisin. J. Volcanol. Geo- cron61ogico do Estado da Bahia, 1 / 1.000.000 - Texto therm. Res. (special issue; in press). explicativo. SGM-SME (Superintend. Geol. Recurs. Deniel, C., Vidal, Ph., Fernandez, A., Le Fort, P. and Peu- Miner.-Secr. Minas Energ.) Bahia, Salvador. cat, J.J., 1987. Isotopic study of the Manaslu granite McReath, I. and Sabat6, P., 1987. Granitoids of the state (Himalaya, Nepal): inferences on the age and source of Bahia, Brazil: A review. Proc. Int. Symp. on Gran- of Himalayan ieucogranite. Contrib. Mineral. Petrol., ites and Associated Mineralizations. Rev. Bras. Geo- 96: 78-92. ci~nc., 17: 404-414. Fernandes, P.C. d'A. and Sabat6, P., 1987. Condiciona- Newton, R.C., 1987. Petrologic aspects of Precambrian mento estrutural das intrus~es proteroz6icas a Molib- granulite facies terrains: bearing on their origins. In: A. denita de Itaberaba e Pedra d'Agua (Bahia-Brasil). 1 ° KriSner (Editor), Proterozoic Lithosphere Evolution. Simp. Nac. Estudos Tect6nicos, Univ. Feder. Bahia, Am. Geophys. Union, Geodyn. Ser., 17:11-26. Salvador, Abstr. Vol., pp. 67-68. Petta, R.A., 1979. Geoquimica do granito de Gameleira e Figueiredo, M.C.H., 1989. Geochemical evolution of suas relag6es com o embasamento na bacia do m6dio eastem Bahia-Brasil: A probable early Proterozoic Rio de Contas, Bahia Brasil. Dissertation, Mestrado, subduction-related magmatic arc. J. South Am. Earth Univ. Feder. Bahia, Salvador (unpublished). Sci., 2(2): 147-154. Pichavant, M., 1984. The effect of boron on liquidus phase Giuliani, G., Fernandez, P.C. d'A and Rudowski, L., 1988. relationships in the system Qz-Ab-Or-H20 at 1 kbar. The Archean and Proterozoic molybdenum minerali- Eos (Trans. Am. Geophys. Union ), 65:298 (abstract). zations of the Bahia State, Brazil: metallogenic impli- Pichavant, M., 1987. Petrogenesis in lithophile-rich mag- 338 P. SABATI~ET AL.

mas. Int. Symp. on Granites and Associated Mineral- Nota explicativa do mapa tem~ltico "Granitog~nese da isation. SGM-SME (Superintend. Geol. Recurs. Bahia" - folha a 1/250.000 Vit6ria de Conquista. Miner.-Secr. Minas Energ.) Bahia, Salvador, Extend. SGM-SME (Superintend. Geol. Recurs. Miner.-Secr. Abstr., Vol., pp. 139-141. Minas Energ. ) Bahia, Salvador, 1 coloured map (sheet Pin, C. and Carme, F., 1987. Sm-Nd isotopic study of 500 printed, leaflet in press). Ma old oceanic crust in the Variscan belt of Western Sato, K., 1982. Programa de data~6es geocronol6gicas do Europe: the Chamrousse ophiolite complex Western Estado da Bahia. Conv. SME (Secr. Minas Energ.) Alps (France). Contrib. Mineral. Petrol., 96: 406-413. Bahia, USP/CPGEO (Univ. S~o Paulo/Cent. Pesqui. Rudowski, L., 1988. P6trologie et g6ochimie des s6ries Geocronol. ), 1978, s.n.t, relat~rios internos). 6volutives de granites peralumineux des massifs de Vidal, Ph., 1987. Use and misuse of radiogenic isotopes Campo Formoso et Carnaiba et leurs relations avec les in granite petrology. Proc. Int. Symp. on Granites and min6ralisations h 6meraudes (Serra de Jacobina, Bahia, Associated Mineralizations. Rev. Bras. Geoci~nc., 17: Br6sil). An. VII Congr. Lat.-Am. Geol. Bel6m, 1: 567- 468-472. 575. Vidal, Ph., Dosso, L., Bowden, P. and Lameyre, J., 1979. Rudowski, L., 1989. P6trologie et g6ochimie des granites Strontium isotope goechemistry in syentie-alkaline transamazoniens de Campo Formoso et de Carnaiba granite complexes. In: A. Ahrens (Editor), Origin and (Bahia, Br6sil) et des phlogopitites h 6meraude asso- Distribution of the Elements. Pergamon, Oxford, pp. ci6es. Doct. Thesis, University of Paris VI, M6m. Sci. 223-231. Terre, Paris, No. 89-15, 282 pp. Vidal, Ph., Cocherie, A. and Le Fort, P., 1982. Geochem- Rudowski, L., Giuliani, G. and Sabat6, P., 1987. Les ical investigations of the origin of the Manaslu leuco- phlogopitites ~ 6meraudes au voisinage des granites de granite (Himalaya, Nepal). Geochim. Cosmochim. Campo Formoso et Carnaiba (Bahia-Br6sil): un ex- Acta, 46: 2279-2292. emple de min6ralisations ~t Be, Mo et W dans des ul- Vidal, Ph., Bernard-Griffiths, J., Cocherie, A., Le Fort, trabasites m6tasomatis6es. C.R. Acad. Sci., Paris, 304, P., Peucat, J.J. and Sheppard, S.M.F., 1984. Geochem- S6r. II, 18:1129-1134. ical comparison between Himalayan and Hercynian Sabat6, P., Machado, G.V.M., da S. and Souza, M.Z.A. leucogranites. Phys. Earth Planet. Inter., 35:179-190. de, 1980. Donn6es structurales des formations pr6- Wilson, N., 1987. Combined Sm-Nd, Pb-Pb and Rb-Sr cambriennes ~pim~tamorphiques du complexe Con- geochronology and isotope geochemistry in polymeta- tendas-Mirante (Bahia, Br6sil). Cah. ORSTOM (Off. morphic precambrian terrains: Examples from Bahia, Rech. Sci. Tech. Outre-mer), ParAs, S6r. G6ol. XI( 1 ): Brazil and Channel Island, U.K.M.Sc. Thesis, Oxford 18-24. University, Oxford (unpublished). Sabat6, P., McReath, I. and Giuliani, G., 1987. Plutonic Wilson, N., Moorbath, S., Taylor, P.N. and Barbosa, J.S., associations of North-East Bahia. Int. Symp. on Gran- 1988. Archean and Early Proterozoic crustal evolution ites and Associated Mineralizations. SGM-SME (Su- in the S~o Francisco Craton, Bahia, Brazil. Eur. Assoc. perintend. Geol. Recurs. Miner.-Secr. Minas Energ.) Geochem., Int. Cong. of Geochemistry and Cosmo- Bahia, Salvador, Excursion Guides Vol., pp. 91-124. chemistry, Paris, Chem. Geol., 70:146 (abstract; spe- Sabat6, P., Gomes, L.C.C. and Anjos, J.A.A., dos, 1988. cial issue).